SA Covid-19 models may grossly overestimate the number of active cases

Piet Streicher PhD Engineering

News24 published a model provided by MediClinic of known active Covid-19 cases per province over the next 9 months. It peaks at 200 000 known active Covid-19 cases for the Western Cape. This scenario is highly unlikely and grossly overestimates Covid-19 cases. 200 000 known active cases in 5.283 million people equates to 34 000 per 1 million people. The worst case countries are the USA at 3 500 / million, UK 3 200 / million and Belgium at 2 800 / million. All three these countries are already well past the peak rate of infections.

Figure 1: Mediclinic model of known active Covid-19 cases by province.

I would like to hear from MediClinic why they expect the Western Cape to peak at a number 10 times higher compared to the worst we have seen in the world? If the MediClinic number of 200 000 only applies to the 17% with health insurance, then I believe they are overestimating the number by a factor of 60+. It is likely that the models do not take account of the number of total infections which could be 10 – 20 times higher than the known infections i.e. the models may have assumed known cases equal all infections. This would be a mistake.

Figure 2: Belgium active cases normalised by population.

Below find the three countries with the highest number of active cases by population:
The USA is at 3 482 active cases / million people.
The UK is at 3 215 active cases / million people.
Belgium is at 2 776 active cases / million people (figure 2).

On a regional basis, New York State has the highest number of active cases at 14 009 / million people. This indicates that the number of active cases by population in a region could significantly exceed the number of active cases by population in a country. The number for New York is 2.5 times lower compared to the number used for the Western Cape in the Mediclinic model.

Figure 3: Belgium new Covid-19 cases by day normalised for population (91-divoc.com).

Belgium, UK and USA are all well past their peak infection rate. The peak infection rate is important as it determines the peak requirement for hospital and ICU beds.

News 24 reports that Mediclinic expects that for 200 000 active cases, they would need 40 000 hospital beds and 10 000 ICU beds. This is a ratio of 20 to 4 to 1. However, we currently have 6 195 active cases in the Western Cape of which 424 are in hospital and 111 are in ICU. This is a ratio of 56 to 4 to 1.

A more realistic projection of the worse case for all hospitals in the Western Cape will be the following:

20 000 active cases in the entire Western Cape when we reach the peak infection rate

1 429 in hospital when we reach peak infection rate

357 in ICU when we reach peak infection rate

If private hospitals only treat the insured and assuming only 17% are insured, that would equate to 61 people in ICU in private hospitals in the Western Cape at the peak. The 357 ICU number is 28 times lower than 10 000 and the 61 ICU number is 164 times lower.

My analysis and projections above are based on the scenario where our active cases are similar to those of the worst performing countries. The reality is that South Africa has significantly slower infection rates compared to Belgium for instance (see figures 3 and 4).

Figure 4: South Africa new Covid-19 cases by day normalised for population (91-divoc.com).

In conclusion, the Mediclinic model published by News24 on 20 May grossly overestimates the peak number of active Covid-19 cases in the Western Cape. My analysis assumes that our mitigation measures are no worse than the worst in the world currently. Initial indications are that our mitigation measures are significantly more effective compared to the worst 25 countries worldwide.

Disclaimer:

My calculations and assumptions may contain errors. This article has not been reviewed by experts. The purpose of the article is to help the media, decision makers and the general public to ask the right questions. If an error is spotted, please comment on the article and I will correct this.

Sources:

https://www.news24.com/SouthAfrica/News/mediclinic-predicts-200-000-covid-19-cases-for-people-with-health-insurance-in-wc-by-early-july-20200520

http://www.91-divoc.com/pages/covid-visualization/

https://mymaskprotectsyou.com/2020/05/14/covid-19-a-call-for-smart-mitigation-measures-considering-the-reduced-risk-for-south-africa-from-covid-19-2/



4 thoughts on “SA Covid-19 models may grossly overestimate the number of active cases

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: